A 50-micrograms/kg dose of fentanyl, given intravenously in divided doses to dogs under enflurane-nitrous oxide anesthesia caused sharp decreases in heart rate (HR), arterial blood pressure (AP), left ventricular dP/dt, and plasma levels of catecholamines. Naloxone, 20 micrograms/kg given 65-70 min later, completely and rapidly reversed these changes. Because the cardiovascular effects of fentanyl and naloxone occurred in unparalyzed animals under surgical anesthesia without eliciting any motor responses, it seems unlikely that they can be ascribed to changes in awareness and surgical stimulation, especially pain. The brief duration of exposure to the narcotic makes it improbable that the naloxone response is due to acute dependence and precipitated withdrawal. Pretreatment with 20 micrograms/kg of atropine only attenuated the decrease in HR, indicating a minor role of vagal mechanisms under these conditions. Administration of 20 micrograms/kg of clonidine by slow infusion after fentanyl further reduced sympathetic activity and greatly attenuated the naloxone response. Injection of 5 mg/kg of tolazoline after administration of clonidine produced massive cardiovascular stimulation by antagonizing clonidine and unmasking the naloxone reversal of fentanyl. Thus, in fully anesthetized dogs, fentanyl decreased the level of cardiovascular function mainly by reducing sympathetic activity. This effect does not seem to be secondary to analgesia or other sensory depressant effects of the narcotic, but rather to an action on central opioid-sensitive mechanisms regulating cardiovascular function.